Anthropic, the company behind the Claude AI assistant, has introduced an updated feature called "projects" in Claude Code, designed to help users manage complex tasks more efficiently. The new system allows users to run multiple AI agents—referred to as "threads"—within the same project. These threads can work on different tasks simultaneously, sharing a common memory, goals, and access to files and other resources. A central "coordinator" oversees the entire process, ensuring tasks are organized and progressing smoothly. Each thread operates as an independent Claude Code session, working on its own copy of the project's repository. If multiple threads need to modify the same part of the code, the system handles it like a typical merge conflict in software development, resolving overlaps in a structured way. This approach helps maintain code integrity even when multiple tasks are happening at once. Additionally, threads can break down their assigned work further using subagents, loops, or predefined workflows, allowing for faster completion of large or complex projects. Users have the flexibility to interact with each thread individually or monitor the overall progress through the main project chat. For now, all threads run in the cloud, but Anthropic has announced that support for running tools and code locally is coming "very soon." This update aims to provide greater flexibility for users who prefer or require local processing. The projects feature is currently in beta and available to a select group of Claude Pro and Max subscribers. Over time, access will expand to include all Pro, Max, Team, and Enterprise users, as well as users of the Cowork and regular Claude chat platforms. This rollout marks a significant step forward in enhancing collaboration and productivity for developers and teams using Claude Code.